Bimodal Control Flow Prediction with State Reversal Circuitry

Solution Overview

Problem

Bimodal control flow prediction circuitry is unable to accurately consider patterns of control flow instruction outcomes, leading to incorrect predictions for control flow instructions that follow specific sequences.

Innovation Solution

Incorporating storage circuitry to store the state of the data processing apparatus associated with control flow instructions and reversal circuitry to reverse predictions based on matching stored and current states, allowing for improved prediction accuracy while maintaining fast and compact prediction capabilities.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Speed

If bimodal control flow prediction circuitry is used, then prediction speed and spatial compactness are improved, but prediction accuracy deteriorates for control flow instructions following specific sequences

Engineering Contradiction:

Improveprediction speedVSAvoidprediction accuracy

Core Design Contradiction:

SpeedVSReliability

Solution Approach 1:

The prediction system is segmented into multiple independent components: bimodal prediction circuitry for base prediction, storage circuitry for state history, and reversal circuitry for correction. Each component handles a specific aspect of prediction, allowing the system to maintain fast bimodal prediction while adding accuracy through separate state tracking and reversal logic.

Inventive Principle:

Principle #1Segmentation

Solution Approach 2:

The system implements feedback by storing the state of the data processing apparatus and using this stored state to reverse predictions when the current state matches the stored state. This feedback mechanism allows the system to learn from previous outcomes and correct prediction errors, improving accuracy without sacrificing the speed of the underlying bimodal predictor.

Inventive Principle:

Principle #23Feedback

2Device complexity

If bimodal control flow prediction circuitry is used, then device complexity is reduced, but prediction accuracy for sequential patterns deteriorates

Engineering Contradiction:

Improveprediction circuit complexityVSAvoidprediction accuracy

Core Design Contradiction:

Device complexityVSReliability

Solution Approach 1:

The prediction system is segmented into multiple independent components: bimodal prediction circuitry for base prediction, storage circuitry for state history, and reversal circuitry for correction. Each component handles a specific aspect of prediction, allowing the system to maintain fast bimodal prediction while adding accuracy through separate state tracking and reversal logic.

Inventive Principle:

Principle #1Segmentation

Solution Approach 2:

Storage circuitry acts as an intermediary between the bimodal predictor and the reversal logic. It stores state information and enables the reversal circuitry to access this history, allowing accurate prediction of sequential patterns without requiring the entire prediction logic to be complex. The intermediary stores and retrieves state data, enabling simple circuitry to achieve accurate predictions.

Inventive Principle:

Principle #24Intermediary (Mediator)

Data Source

PatentUS11915004B2Control flow prediction

Publication Date: 2024.02.27 ARM LTD

AI summary

A data processing apparatus is provided that includes bimodal control flow prediction circuitry for performing a prediction of whether a conditional control flow instruction will be taken. Storage circuitry stores, in association with the control flow instruction, a stored state of the data processing apparatus and reversal circuitry reverses the prediction in dependence on the stored state of the data processing apparatus corresponding with a current state of the data processing apparatus when execution of the control flow instruction is to be performed.